A computer-assisted pproach to the comparison of mainland southeast Asian languages

This cumulative thesis is based on three separate projects based on a computer-assisted language comparison (CALC) framework to address common obstacles to studying the history of Mainland Southeast Asian (MSEA) languages, such as sparse and non-standardized lexical data, as well as an inadequate method of cognate judgments, and to provide caveats to scholars who will use Bayesian phylogenetic analysis. The first project provides a format that standardizes the sound inventories, regulates language labels, and clarifies lexical items. This standardized format allows us to merge various forms of raw data. The format also summarizes information to assist linguists in researching the relatedness among words and inferring relationships among languages. The second project focuses on increasing the transparency of lexical data and cognate judg- ments with regard to compound words. The method enables the annotation of each part of a word with semantic meanings and syntactic features. In addition, four different conversion methods were developed to convert morpheme cognates into word cognates for input into the Bayesian phylogenetic analysis. The third project applies the methods used in the first project to create a workflow by merging linguistic data sets and inferring a language tree using a Bayesian phylogenetic algorithm. Further- more, the project addresses the importance of integrating cross-disciplinary studies into historical linguistic research. Finally, the methods we proposed for managing lexical data for MSEA languages are discussed and summarized in six perspectives. The work can be seen as a milestone in reconstructing human prehistory in an area that has high linguistic and cultural diversity.